Applied Computing Raises $20M to Put a Foundation Model Inside Entire Oil and Gas Plants

The London startup's Orbital model fuses time-series data, physics equations and language into one system that can flag an anomaly, chase its root cause and simulate whether the fix breaks something else — with engineering giant KBR leading the round.

Jul 16, 2026Source: TechCrunch

Applied Computing Raises $20M to Put a Foundation Model Inside Entire Oil and Gas Plants

Industrial AI usually means a model that watches one machine. Applied Computing is selling a model that claims to understand the whole plant — and on July 16 it disclosed a $20 million Series A led by engineering and government-services group KBR, with Databricks Ventures participating, to push the idea into refineries and petrochemical sites.

What Orbital actually is

The company's model, Orbital, is a physics-informed multi-modal foundation model — not a language model with an industrial skin. It fuses three components in real time: a time-series model that connects to distributed control systems, data historians and lab information systems to forecast trends and catch anomalies; a physics model that extracts governing equations from equipment manuals and enforces mass balance, energy conservation and reaction kinetics; and a language model trained on chemical engineering knowledge that reads P&IDs, standard operating procedures and work orders. The pitch is that Orbital can flag an anomaly, investigate its root cause, then simulate whether a proposed fix creates a problem elsewhere — compressing multi-day investigations into minutes.

The founders

Applied Computing was founded in 2023 and is an Imperial College London-linked spinout, with headquarters in London, an operational hub in Bengaluru and now a Houston office funded by this round. Co-founder and CEO Callum Adamson was an expert-in-residence at Imperial's Enterprise Lab, where he mentored his eventual co-founder. Chief AI Officer Dr. Samyakh Tukra holds a PhD in AI and 3D computer vision from Imperial's Hamlyn Centre for Robotic Surgery and previously worked at Shell and Tractable. President Dan Jeavons, hired in 2025, ran Shell's AI program.

Traction, and a circular customer

The company says it moved from stealth to double-digit millions in ARR in under 18 months, a figure it has not broken out. KBR has integrated Orbital into its INSITE 3.0 platform, launched in March, and is applying it to ammonia production — making KBR simultaneously lead investor, customer and channel partner. Wipro and Databricks signed on in March as deployment partners for the Middle East, India and Southeast Asia. Other customers are unnamed, including what the company describes as a major US upstream operator.

The competition

Orbital lands against AspenTech and AVEVA in physics-based process simulation and Cognite and Seeq in industrial data tooling. Adamson's framing is that rivals bolt AI onto a simulator: "It's getting those three data sources to talk to each other in real time. That's the real key," he told TechCrunch. He claims facilities today make operating decisions on less than 8% of the data they collect.

Why It Matters

Heavy industry is the last big AI market where hallucination is not an inconvenience but a safety incident, which is exactly why a model constrained by conservation laws is a more plausible sell than a chatbot. The tell to watch is KBR's triple role as investor, customer and reseller — it derisks distribution but means the revenue proof point is not yet an arms-length one. If physics-informed models work at plant scale, the incumbent simulation vendors face their first real architectural challenge in decades.

Filed underAI Models Manufacturing AI #Applied Computing #KBR #foundation models #industrial AI #oil and gas